Genomic Deletion of Chromosome 12P Is an Independent Prognostic Marker in Prostate Cancer

www.impactjournals.com/oncotarget/ Oncotarget, Vol. 6, No. 29

Genomic deletion of chromosome 12p is an independent prognostic marker in prostate cancer

Martina Kluth1, Ramin Ahrary1, Claudia Hube-Magg1, Malik Ahmed1, Heinke Volta1, Catina Schwemin1, Stefan Steurer1, Corinna Wittmer1, Waldemar Wilczak1, Eike Burandt1, Till Krech1, Meike Adam2, Uwe Michl2, Hans Heinzer2, Georg Salomon2, Markus Graefen2, Christina Koop1, Sarah Minner1, Ronald Simon1, Guido Sauter1, Thorsten Schlomm2,3 1Institute of Pathology, University Medical Center Hamburg-Eppendorf, Germany 2Martini-Klinik, Prostate Cancer Center, University Medical Center Hamburg-Eppendorf, Germany 3Dept. of Urology, Section for translational prostate cancer research, University Medical Center Hamburg-Eppendorf, Germany Correspondence to: Ronald Simon, e-mail: [email protected] Keywords: 12p deletion, prostate cancer, CDKN1B, p27, prognostic marker Received: May 27, 2015 Accepted: July 03, 2015 Published: July 21, 2015

ABSTRACT

Deletion of 12p is a recurrent alteration in prostate cancer, but the prevalence and clinical consequences of this alteration have not been studied in detail. Dual labeling fluorescence in situ hybridization using probes for 12p13 (CDKN1B; p27) and centromere 12 as a reference was used to successfully analyze more than 3700 prostate cancers with clinical follow-up data assembled in a tissue microarray format. CDKN1B was selected as a probe because it is located in the center of the deletion, which spans > 10 Mb and includes > 50 genes in 80% of cancers with 12p deletion. Deletion of 12p was found in 13.7% of cancers and included 13.5% heterozygous and 0.2% homozygous deletions. 12p deletion were linked to advanced tumor stage (p < 0.0001), high Gleason grade (p < 0.0001), rapid tumor cell proliferation (p < 0.0001), lymph node metastasis (p = 0.0004), and biochemical recurrence (p = 0.0027). Multivariate analysis including pT stage (p < 0.0001), Gleason grade (p < 0.0001), pN status (p = 0.0001), preoperative PSA levels (p = 0.0001), and resection margin status (p = 0.0001) revealed an independent prognostic value of 12p deletion (p = 0.0014). Deletion of 12p was unrelated to the ERG fusion status. Deletion of 12p was only marginally linked to reduced p27 expression, which by itself was unrelated to clinical outcome. This argues against p27 as the key target gene of 12p deletions. In summary, the results of our study demonstrate that 12p deletion is frequent in prostate cancer and provides independent prognostic information. 12p deletion analysis alone, or in combination with other prognostic parameters may thus have clinical utility.

INTRODUCTION will develop life-threatening disease that requires radical treatment [2, 3]. As screening strategies identify prostate Prostate cancer is the most frequent cancer in males. cancers already at early stages of the disease, it becomes About 900,000 men are diagnosed with this disease every increasingly important to avoid overtreatment of patients year, and almost 130,000 eventually die from their cancer with less aggressive disease. Establishing molecular in Western societies alone [1]. Autopsy studies suggest markers enabling a distinction between indolent and that more than 50% of males will develop prostate cancer aggressive forms of the prostate cancer is thus of utmost during their lifetime, but only a minority of affected men importance.

www.impactjournals.com/oncotarget 27966 Oncotarget Chromosomal rearrangements are of considerable clinical and pathological parameters in the 3,757 cancers interest as diagnostic or prognostic biomarkers amongst with interpretable FISH results and the 3,725 cancers others because they can reproducibly be tested due to without interpretable FISH results was comparable. their dichotomous nature being either present or not. Structural chromosomal rearrangements represent a major 12p deletions and prostate cancer phenotype mechanism for activating and inactivating oncogenes and tumor suppressor genes in cancer. In contrast to most 12p deletions were found in 13.7% (514 of 3,757) other malignancies, chromosomal rearrangements only of all prostate cancers, including 13.5% heterozygous and rarely involve copy number gains or amplifications in 0.2% homozygous deletions. The relationship between prostate cancer. With the exception of the TMPRSS2:ERG 12p deletions and tumor phenotype and clinical parameters fusion affecting about 50% of prostate cancers, all other is summarized in Table 1. 12p deletions were significantly individual translocations also occur at very low frequency linked to high Gleason grade (p < 0.0001), advanced tumor (<5%) [4–6]. Many chromosomal deletions, however, stage (p < 0.0001), presence of lymph node metastasis are highly recurrent and occur in > 10% of cancers. The (p = 0.0004), and elevated preoperative PSA values most common deletions include 8p (40–50%), 13q14, (p = 0.0027). 16q22-q24, 6q12-q22, 10q23 (20–30% each), 12p12-p13, 3p13 (15–20% each), and 5q21 (10%) [6–10]. It is not 12p deletion and ERG fusion status fully understood, how these deletions impact prostate 12p deletions were unrelated to ERG fusion status cancer cells and their exact mechanisms of action may irrespective from the method of ERG analysis (p = 0.5626 vary between deleted loci. Very small deletions may for ERG-IHC and p = 0.9790 for ERG-FISH analysis). impact one specific gene. For example deletions at 10q23 Deletions of 12p were found in 14.4% and 16.0% of are typically narrow and are likely to specifically target ERG-negative cancers (according to ERG IHC and FISH PTEN, a gene with tumor suppressive properties, which analysis), and in 13.7% (IHC, p = 0.5626) and 16.0% has been widely studied in prostate cancer [6, 8, 11]. Most (FISH, p = 0.9790) of ERG-positive cancers (Figure 1). other deletions are substantially larger and may include There was no major difference in the relationship between hundreds of genes [6–10]. Even though some of these 12p deletions and tumor phenotype between ERG-positive deletions contain known tumor suppressor genes, such as and ERG-negative cancers. Most associations of 12p the retinoblastoma gene on 13q14, it remains doubtful that deletions and tumor phenotype parameters held also true the loss of large pieces of DNA just serves the purpose in subgroup analyses (Table 1). of affecting just one single gene. The one deletion / one gene hypothesis is also disturbed by the adamant lack of 12p deletion and p27 expression inactivating mutations seen in the retained allele of the best candidate tumor suppressor genes in studies utilizing p27 expression data were available from 2,125 next generation sequencing [4, 5, 12, 13]. patients for whom 12p deletion data were also available. Irrespective of their impact on gene function, several p27 was negative in 16.6%, weak in 34.8%, moderate in deletions have shown a striking prognostic impact in 29.4%, and strong in 19.2% of these cases. Loss of p27 prostate cancer. This applies for example for deletions of immunostaining was linked to tumors of low Gleason PTEN [11], 17p13 [14], 5q21 [15], 6q15 [16], and 8p [17]. grade (P < 0.0001) and ERG fusion negative cancers Other frequent deletions, such as deletions of 12p, have (P < 0.0001). Reduced (negative or weak) p27 expression so far not been analyzed for their potential prognostic was found in 59.7% of 12p deleted and in 50.4% of 12p role. In order to study the prognostic impact of 12p copy undeleted cancers (p = 0.0080; Table 2). number alterations, we analyzed more than 7,000 prostate cancers with clinical follow-up data by fluorescencein situ 12p deletions and clinical outcome hybridization (FISH). The results of our study identify 12p deletion as a strong independent molecular prognostic Follow-up data were available from 3,521 tumors feature in prostate cancer. that were successfully analyzed for 12p deletion. In univariate analysis, 12p deletions were strongly linked RESULTS to early biochemical (PSA) recurrence in all cancers (p < 0.0001, Figure 2a) and there was no difference Technical aspects seen in the prognostic impact of 12p deletions 1, 701 ERG-negative (p < 0.0001, Figure 2b) and 1,578 ERG- 12p FISH analysis was successful in 3,757 of 7,482 positive cancers (p < 0.0001, Figure 2c). Significant (50.2%) arrayed cancers. Analysis was not informative in associations with PSA recurrence held also true in the remaining 3,725 tumors because of lack of tumor cells subsets of cancers with low (Gleason ≤ 3 + 4, p = 0.0003) in the tissue spots, faint or absent FISH signals, or missing or high Gleason (≥ 4 + 3, p < 0.0001, Figure 2d–2e). tissue spots on the TMA section. The distribution of 12p deletion status was also significantly associated www.impactjournals.com/oncotarget 27967 Oncotarget Table 1: Associations between 12p deletion and prostate cancer phenotype in all, ERG fusion positive, and ERG fusion negative tumors All cancers ERG-negative cancers ERG-positive cancers n 12p p value n 12p p value n 12p p value deletion deletion deletion (%) (%) (%) 3757 13.7% 1821 14.4% 1678 13.7% Tumor stage pT2 2312 11.4% <0.0001 1176 12.2% <0.0001 960 11.8% 0.0630 pT3a 911 15.5% 397 13.9% 468 16.9% pT3b 468 20.5% 219 26.5% 218 14.7% pT4 42 21.4% 21 23.8% 17 17.7% Gleason grade ≤ 3 + 3 1135 8.6% <0.0001 547 7.1% <0.0001 486 10.7% 0.0038 3 + 4 1853 14.5% 861 16.0% 881 13.6% 4 + 3 583 18.5% 303 21.1% 244 16.4% ≥ 4 + 4 161 21.7% 101 19.8% 52 28.9% Lymph node metastasis N0 2160 15.1% 0.0004 1049 15.4% 0.0023 987 15.3% 0.0743 N+ 173 26.1% 82 29.3% 82 23.2% PSA Level (ng/μl) <4 466 10.1% 0.0027 201 11.9% 0.0909 226 8.4% 0.0206 4–10 2128 13.1% 1014 13.4% 967 13.4% 10–20 804 15.6% 438 16.2% 319 16.0% >20 294 18.7% 144 20.1% 130 18.5% Surgical margin negative 2894 13.1% 0.1224 1432 13.8% 0.1831 1251 13.1% 0.6165 positive 796 15.2% 361 16.6% 390 14.1%

to unfavorable outcome according to our alternative stage, resection margin status and preoperative clinical endpoints, which represent the four hallmarks PSA level, 12p deletion predicted early PSA recurrence (local, local invasive, occult systemic and metastatic) independently from these parameters in all cancers of tumor growth and dissemination in prostate cancer (p = 0.0014) and in the subset of ERG-positive cancers (Figure 3). More than 70% of patients without 12p (p = 0.0022). It only marginally fell short of reaching deletion had either local or local invasive disease while statistical significance in ERG-negative cancers this fraction decreased to 53% in patients with 12p p = 0.0723, Table 3). deletion (p < 0.0001, Figure 3a). A separate analysis of low grade (Gleason ≤ 3 + 4 = 7) and high grade (Gleason 12p deletion and cell proliferation ≥ 4 + 3 = 7) cancers revealed that the prognostic impact of 12p deletions was particularly strong in low-grade 12p deletions were linked to increased cell cancers (p = 0.0029) while it largely disappeared proliferation as measured by Ki67 immunohistochemistry. in high grade tumors (p = 0.1032, Figure 3b–3c). This was true for all tumors but also for most subgroups In a multivariate analysis including the established of Gleason ≤ 3 + 3 = 6, 3 + 4 = 7, 4 + 3 = 7, and prognostic predictors pT stage, Gleason grade, nodal ≥ 4 + 4 = 8 (Table 4).

www.impactjournals.com/oncotarget 27968 Oncotarget Figure 1: Relationship between 12p deletion and ERG fusion probed by IHC and FISH.

Table 2: Associations between p27 expression and Gleason grade, ERG fusion status, and 12p deletion p27 status in all cancers n negative weak moderate strong p value Gleason grade ≤ 3 + 3 1305 25.13 32.11 26.28 16.48 <0.0001 3 + 4 1755 15.16 34.64 29.23 20.97 4 + 3 444 15.09 33.78 30.18 20.95 ≥ 4 + 4 145 8.28 33.1 34.48 24.14 ERG fusion status negativ 1602 27.53 38.45 22.28 11.74 <0.0001 positive 1536 6.9 30.92 35.68 26.5 12p deletion status normal 1892 17.6 42.06 28.33 28.33 0.0075 deleted 233 16.49 33.93 29.49 20.08

Location and extension of 12p deletions in DISCUSSION prostate cancer The results of our study demonstrate that 12p Data on the patterns of 12p deletions from own deletion occurs in a relevant fraction of prostate cancers previous analyses and from the literature are summarized and has a strong and independent prognostic relevance in in Figure 4. The data demonstrate that >80% of 12p this tumor. deletions are larger than 10 Megabases (Mb) and contain About 14% of more than 3,500 prostate cancers more than 50 genes. The minimal common deletion is successfully analyzed by FISH showed 12p deletions at 12p13.1, spans about 1.5 Mb and contains 7 genes in our study. Earlier studies employing classical or including DUSP16, CREBL2, GPR19, CDKN1B, array-based comparative genomic hybridization or loss APOLD1, DDX47, and GPRC5A. of heterozygosity (LOH) analysis have often found

www.impactjournals.com/oncotarget 27969 Oncotarget Figure 2: Association of 12p deletion with biochemical recurrence in a. all prostate cancers, b. ERG-positive prostate cancers, c. ERG-negative prostate cancers, d. low-grade cancers, and e. high-grade cancers.

higher rates. Classical CGH studies have described 12p selected threshold of 60% of cells required to have fewer deletions in 17% of 6 [18] and 12% of 52 [19] cancers. 12p than centromere 12 signals to call a “12p deletion” Array based CGH analyses have found 12p deletions in has been earlier extensively validated by our group for 24% of 46 [20], 22% of 48 [21], 27% of 64 [8] and 12% PTEN deletions [11]. of 181 [6] prostate cancers. The somewhat higher rate Deletions of 12p were significantly associated of 12p deletions in many of these studies may partly be with advanced tumor stage, high Gleason grade, high due to a bias caused by small patient numbers or patient preoperative PSA value, lymph node metastasis, selection, as several projects focused on advanced, accelerated tumor cell proliferation, and early biochemical hormone refractory or metastatic cancers [21–24]. The recurrence. These findings are in line with earlier data, highest rates of 12p deletions described in prostate cancer which already suggested a possible link of 12p deletions are derived from PCR-based loss of heterozygosity with unfavorable tumor phenotype. Taylor et al. described (LOH) studies. Here, authors described 12p LOH in 12% a significant association of 12p deletion with Gleason of 17 [24], 23% of 99 [22] and 49% of 19 cancers [23]. grade and pT stage in a series of 181 cancers [6]. The The high rates of LOH found in some studies may be mechanism causing increased aggressiveness of cancer due to some false positive cases caused by chromosome cells with 12p deletion is not clear. Own earlier data and 12 polysomy in non-diploid tumors that account for 30– data from the literature demonstrate, that the deleted area 50% of prostate cancers [25, 26]. FISH is regarded as typically spans more than 25 Mb (Figure 4) and contains the gold standard for gene copy number determination as more than 100 genes. We selected a FISH probe targeting it allows for analysis on the single cell level, and is not the CDKN1B locus (Cyclin-Dependent Kinase Inhibitor disturbed by contaminating non-neoplastic cells that are 1B, 12p13.1) because CDKN1B is located in the center inevitably present in cancer tissues. While other FISH of the commonly deleted 12p region [6], and it encodes studies dealing with 12p deletions in prostate cancer are p27/kip1, a well known tumor suppressor gene with lacking, we are confident with our data, also because the importance in various cancer types (reviewed in [27, 28].

www.impactjournals.com/oncotarget 27970 Oncotarget Figure 3: Association of 12p deletion with clinical groups representing the clinical hallmarks of prostate cancer in a. all tumors and subsets of b. low-grade and c. high-grade cancers.

Several studies have utilized IHC to analyze p27 in outside the typical deletion regions (e.g. SPOP at prostate cancer and reported a reduced p27 expression 17q21, FOXA1 at 14q12, or ZNF595 at 4p16), and in 12.5%–84.3% of tumors. Several authors described for the common large deletions such as 8p, 5q, 6q, that reduced p27 expression was linked to unfavorable 16q, and 13q, recurrent mutation have not been found tumor phenotype and prognosis [29–38] but others [4–6, 12, 13]. These findings challenge the classical could not confirm these observations [33, 34, 39–43]. recessive model of biallelic tumor suppressor gene In an own study on 4,699 carcinomas, we also failed inactivation. The biological mechanism driving the to see a prognostic impact of reduced p27 expression development of large deletions involving hundreds of [44]. Altogether the published data argue against a genes is not known. It is tempting to speculate, however, strong prognostic impact of reduced p27 expression in that simultaneous dosage reduction of multiple genes prostate cancer. The only marginal association of reduced within such a deletion jointly support tumor growth. p27 expression with 12p deletion in this study further Such cooperative effects have indeed reported from argues against p27 playing a critical role for increased mouse orthologs of human 8p11-p23 genes in a mouse aggressiveness of 12p deleted cancer cells. model of HCC [45]. Of note, large 8p deletion, often The one deletion / one gene hypothesis involving the entire chromosome arm, is one of the most is currently challenged by the adamant lack of frequent alterations in many solid tumor types including inactivating mutations seen in the retained allele of prostate cancer and attempts to identify “the critical 8p the best candidate tumor suppressor genes in studies gene” have so far failed [6–8]. Beyond the cell cycle utilizing next generation sequencing [4, 5]. In fact, regulator CDKN1B (12p13.1), further potentially the most recurrently mutated genes are located relevant cancer genes on 12p may for example include

www.impactjournals.com/oncotarget 27971 Oncotarget Table 3: Multivariate analysis including established prognostic parameters and 12p deletion in all, ERG fusion positive, and ERG fusion negative tumors all prostate cancers ERG-negative prostate cancers ERG-positive prostate cancers Parameter RR 95% CI p-value RR 95% CI p-value RR 95% CI p-value Tumor stage pT3a vs 2.0 1.6–2.5 <0.0001 1.8 1.4–2.4 <0.0001 2.1 1.5–2.8 <0.0001 pT2 pT3b vs 1.6 1.3–1.9 1.3 1.0–1.8 2.1 1.6–2.7 pT3a pT4 vs 1.8 1.2–2.5 1.9 1.1–3.1 1.1 0.6–1.9 pT3b Gleason grade 3 + 4 vs 2.0 1.5–2.6 <0.0001 1.5 1.1–2.3 <0.0001 2.9 1.9–4.6 <0.0001 ≤ 3 + 3 4 + 3 vs 2.1 1.8–2.5 2.0 1.6–2.6 2.1 1.6–2.7 3 + 4 ≥ 4 + 4 vs 1.0 0.8–1.3 1.0 0.7–1.4 1.2 0.8–1.8 4 + 3 Nodal stage pN1 vs 1.6 1.3–2.0 0.0001 1.8 1.2–2.5 0.0024 1.3 0.9–1.8 0.1351 pN0 Resection margin status R1 vs R0 1.4 1.2–1.6 0.0001 1.5 1.2–1.9 0.0010 1.3 1.0–1.6 0.0340 Pre-operative PSA (ng/ml) 4–10 vs <4 1.0 0.7–1.3 <0.0001 1.0 0.6–1.6 0.0005 0.9 0.6–1.4 0.1668 10–20 vs 1.3 1.1–1.5 1.5 1.2–1.9 1.1 0.9–1.4 4–10 >20 vs 1.3 1.0–1.6 1.2 0.9–1.6 1.3 0.9–1.7 10–20 12p status 12p deletion vs 1.4 1.1–1.6 0.0014 1.3 1.0–1.7 0.0723 1.6 1.2–2.1 0.0022 normal

CD9, ING4, and BCL-G. CD9 (12p13.31) regulates development of certain deletions or vice versa, and that multiple cellular functions such as cell adhesion, certain deletions may induce gene expression changes migration, apoptosis, and tumor cell motility [46], ING4 that inhibit the development of ERG rearrangements (12p13.31) is involved in cell cycle arrest, apoptosis and [4–6, 48]. Deletions that are independent of the ERG senescence [47], and BCL-G (12p13.2) plays a role in status have so far not been described. The independence of apoptosis [47]. 12p deletions of ERG rearrangements is thus remarkable Chromosomal deletions analyzed in large prostate and might suggest, that either different – ERG independent cancer cohorts have so far shown striking associations – mechanisms may apply for 12p deletion development or with either ERG-positive or ERG-negative cancers. For that – in contrast to most other recurrent deletions - none example, deletions of 6q15 and 5q21 are frequent in ERG of the affected genes is influencing possible mechanisms fusion negative cancers, whereas deletions of 3p13, TP53 needed for the development of TMPRSS2:ERG fusion. and PTEN are common in ERG fusion positive cancers. The strong link of 12p deletions with the probability It has been suggested, that ERG induced “reprogramming” for PSA recurrence is of potential clinical relevance of the cellular environment can facilitate or inhibit the especially since this was independent of established clinical www.impactjournals.com/oncotarget 27972 Oncotarget Table 4: Association between cell proliferation as measured by Ki67LI immunohistochemistry and 12p deletion in tumor subsets of different Gleason grade 12p status n evaluable Ki67 Li mean Std.deviation all cancers P < 0.0001 normal 2220 2.67 0.05 deleted 385 3.31 0.13 pGleason ≤ 3 + 3 P = 0.0172 normal 696 2.22 0.08 deleted 75 2.85 0.25 pGleason 3 + 4 P = 0.0017 normal 1118 2.67 0.07 deleted 202 3.20 0.16 pGleason 4 + 3 P = 0.8961 normal 311 3.26 0.18 deleted 81 3.31 0.35 pGleason ≥ 4 + 4 P = 0.1025 normal 81 4.10 0.48 deleted 24 5.75 0.88

Figure 4: Size and extension of chromosome 12p deletions detected in published microarray-based copy number studies [6, 55–59]. Each bar represents the deleted area in a single tumor. The minimal commonly deleted region is indicated be red dotted lines. www.impactjournals.com/oncotarget 27973 Oncotarget and pathological parameters. However, the selection of undergoing surgery at the Department of Urology, and the clinical endpoints is not trivial in studies evaluating prostate Martini Clinic, Prostate Cancer Center, University Medical cancer. It is obvious, that PSA recurrence does not represent Center Hamburg-Eppendorf. This TMA is based on our an optimal clinical endpoint, even though it is used in the previous 3,261 samples prostate prognosis TMA [17], with vast majority of studies. More than 30% of prostate cancer additional 4,634 tumors and updated clinical data from 6,894 patients experience a PSA recurrence after surgical therapy patients with a median follow-up of 38.8 months (range: within ten years, but only a minority of these cancers 1 to 241 months; Table 5). In all patients, prostate specific (<10%) will eventually progress to life threatening disease antigen (PSA) values were measured quarterly in the first [49]. Moreover, PSA recurrence barely represents the year, followed by biannual measurements in the second and natural history of prostate cancer after radical prostatectomy. annual measurements after the third year following surgery. Even cancer specific death – often considered as an optimal Different clinical endpoints were defined in the patient set: clinical endpoint - is suboptimal in prostatectomy cohorts Time to recurrence was defined as the time interval between because knowing the risk to die from one’s cancer despite surgery and first occurrence of a postoperative PSA of maximal multimodal therapy concepts - including radical ≥0.2 ng/ml and rising thereafter. Patients without evidence surgery, adjuvant radiation and/or hormonal- and cytotoxic of tumor recurrence were censored at the time of the last therapy - has limited practical value in the absence of follow-up. In addition, patients were grouped into subsets established additional adjuvant treatment options. defining the milestones of tumor progression: Group 1 In an attempt to define clinical endpoints that “organ confined tumor growth” included 1,458 patients with could better support pre-therapeutic decision making organ confined tumors (pT2) and no biochemical relapse at the time of the initial cancer diagnosis we defined (BCR) in long-term follow-up (at least 3y after surgery; clinical groups, representing the two clinical hallmarks mean 122.2 (120.5–129.9) months), Group 2 “local invasive of cancer, i.e., local tumor extension and systemic tumor tumor growth” included 446 patients with histological proof growth and that can be analyzed on materials derived of extraprostatic tumor growth (pT3a or pT3b), but no BCR from prostatectomy specimen. These clinical groups were in long-term follow-up (at least 3y after surgery; mean selected to reflect the “biological milestones” of cancer 125.2 (121.2 – 129.2) months) or pT3 tumors with BCR progression, including 1) organ-confined tumors (pT2) but permanent response to secondary adjuvant or salvage without relapse in long-term follow-up – and thus assumed radiation therapy, Group 3 “occult systemic dissemination” “local”, 2) “local invasive” cancers (pT3) without relapse included 517 cancers with biochemical failure after two or with permanent response to local secondary radiation local therapies (radical prostatectomy and additional in long-term follow-up, 3) “occult systemic” disease adjuvant or salvage- radiation therapy), and without characterized by failure of two local therapies (radical development of distant metastases in long-term follow-up prostatectomy and secondary radiation), but no evidence (mean follow-up 87.1 (83.3–91.0) months), and Group 4 of distant metastases in long-term follow-up, and 4) “metastatic tumor growth” included 199 patients with “metastatic” disease characterized by regional or distant development of distant (bone and/or visceral) or lymph node metastases. That this analysis also revealed a strikingly metastases (mean follow-up 95.5 (89.0–102.0) months). increased risk for presence of occult systemic or manifest In group 1–3, patients with neoadjuvant or post operative metastatic disease in 12p deleted further supports the hormonal therapies were excluded. The remaining 4 714 potential clinical significance of this deletion. cancers could not be classified into one of these groups, In summary, these data identify 12p deletion as a mainly because clinical information was incomplete or the frequent event in prostate cancer, which is unrelated to follow-up time was to short. All prostate specimens were the ERG fusion status but strongly linked to aggressive histologically examined according to a standard procedure, tumor behavior. The large average size of the deletion, the including complete embedding of the entire prostate for only marginal association with p27 protein expression and histological analysis [50]. The TMA manufacturing process the lack of prognostic impact of p27 expression argues was described earlier in detail [51, 52]. In short, one against p27 representing the sole or critical target gene. 0.6 mm core was taken from a representative tissue block Irrespective of the target gene(s), 12p deletion analysis from each patient. The tissues were distributed among may have clinical utility for prognostic testing of prostate 16 TMA blocks, each containing 144 to 522 tumor samples. biopsies either alone or in combination with other features. Presence or absence of cancer tissue was validated by immunohistochemical AMACR and 34BE12 analysis on MATERIALS AND METHODS adjacent TMA sections. Each TMA block also contained various control tissues, including normal prostate tissue. Patients Analysis of patient and corresponding histopathological data for research purposes, as well as construction of tissue A set of prostate cancer tissue microarrays (TMA) was microarrays from archived diagnostic left-over tissues, was used in this study containing one tissue core each from 7,482 approved by local laws (HmbKHG, §12, 1) and by the local consecutive radical prostatectomy specimens from patients ethics committee (Ethics commission Hamburg, WF-049/09 www.impactjournals.com/oncotarget 27974 Oncotarget Table 5: Clinico-pathological features of 7 482 arrayed prostate cancers No. of patients (%) Study cohort on TMA (n = 7482) Biochemical relapse among categories Follow-up (mo) n 7482 (60.2%) 1457 (19.5%) Mean 53.4 - Median 36.8 - Age (y) ≤50 234 (3.2%) 43 (18.4%) 51–59 1912 (25.8%) 368 (19.2%) 60–69 4438 (59.9%) 872 (19.6%) ≥70 822 (11.1%) 172 (20.9%) Pretreatment PSA (ng/ml) <4 976 (13.2%) 125 (12.8%) 4–10 4443 (60.3%) 650 (14.6%) 10–20 1461 (19.8%) 411 (28.1%) >20 488 (6.6%) 248 (50.8%) pT category (AJCC 2002) pT2 4927 (66.2%) 460 (9.3%) pT3a 1650 (22.2%) 477 (28.9%) pT3b 803 (10.8%) 472 (58.8%) pT4 58 (0.8%) 48 (82.8%) Gleason grade ≤ 3 + 3 2316 (31.2%) 171 (7.4%) 3 + 4 3804 (51.2%) 693 (18.2%) 4 + 3 1018 (13.7%) 448 (44%) ≥ 4 + 4 287 (3.9%) 144 (50.2%) pN category pN0 3963 (92.4%) 919 (23.2%) pN+ 328 (7.6%) 207 (63.1%) Surgical margin Negative 5921 (80.6%) 914 (15.4%) Positive 1428 (19.4%) 516 (36.1%)

NOTE: Numbers do not always add up to 7482 in the different categories because of cases with missing data. Abbreviation: AJCC, American Joint Committee on Cancer. and PV3652). All work was carried out in compliance with Fluorescence in-situ hybridization the Helsinki Declaration. The molecular database attached to this TMA includes data on ERG expression in 6,135, on Four micrometer TMA sections were used for ERG rearrangement by FISH analysis in 3,688 (extended fluorescence in-situ hybridization (FISH). TMA sections from [53]), as well as immunohistochemical data on p27 were de-waxed, air-dried, and dehydrated in 70%, 85%, expression from 3,701 [44] and KI67 labeling Index and 100% ethanol. Slides were pretreated in VP 2000 (Ki67 LI) data in 4,426 patients (extended from [54]). Pretreatment Reagent (Abbott, Des Plaines, USA) for www.impactjournals.com/oncotarget 27975 Oncotarget Figure 5: Examples of FISH findings showing a. normal 12p signal numbers and b. heterozygous 12p deletion. Green signals correspond to centromere 12, red signals correspond to the CDKN1B (12p13) gene locus.

15 min at 80°C, followed by 150 min incubation at Statistics 37°C in 0.5% protease 1 solution (Abbott, Des Plaines, USA). 4 μl of FISH probe mix in 70% formamide 2x For statistical analysis, the JMP software (SAS SSC solution was applied to the slides and co-denatured Institute Inc., NC, USA) was used. Contingency tables with the cellular DNA in a Hybrite hybridization oven were calculated to study association between 12p deletion for 10 min at 72°C prior to overnight-hybridization and clinico-pathological parameters, and the Chi-square at 37°C in a humidified chamber. The FISH probe mix (Likelihood) test was used to find significant relationships. consisted of a spectrum-orange labeled 12p (CDKN1B Kaplan Meier plots were generated for PSA recurrence-free locus, 12p13.1) probe (made from BACs RP11–180M15 survival. The Log-Rank test was applied to determine the and BAC RP11–113I21), and a spectrum-green labeled, significance of differences between the survival curves. Cox commercial centromere 12 probe (#6J37–12; Abbott, proportional hazards regression analysis was performed to Wiesbaden, Germany) as a reference. After hybridization, test the statistical independence and significance between slides were subjected to serial stringent washings pathological, molecular, and clinical variables. (2x SSC solution with 0, 3% NP40 at 72°C for 2 minutes) and counterstained with 0.2 μmol/L 4′-6-diamidino-2- ACKNOWLEDGMENTS phenylindole (DAPI) in antifade solution. Stained slides were manually interpreted under an epifluorescence We are grateful to Janett Lütgens, Sünje Seekamp, microscope, and the predominant green and orange Inge Brandt, Bianca Kelp, Anne Meyer and Lisa Paustian FISH signal numbers were recorded in each tissue spot. for excellent technical assistance. Homozygous deletion of 12p was defined as complete lack of 12p FISH signals in the tumor nuclei, but presence of 12p FISH signals in adjacent normal cells. Tissue FUNDINGS spots lacking 12p signals in all (tumor and normal cells), or lacking of any normal cells as an internal control for We certify that there are no sources of any support successful hybridization of the 12p probe, were excluded for the work, received in the form of grants, equipment, from analysis. Heterozygous deletion of 12p was defined and/or drugs in relation to this article. as the presence of fewer 12p signals than centromere 12 probe signals of ≥ 60% tumor nuclei. These thresholds CONFLICTS OF INTEREST were based on a previous study analyzing PTEN deletions in a subset of slides of the TMA set [11]. Representative We certify that there is no actual or potential conflict FISH images are shown in Figure 5. of interest in relation to this article. www.impactjournals.com/oncotarget 27976 Oncotarget REFERENCES in ERG fusion-positive and fusion-negative prostate cancer. The American journal of pathology. 2012; 181:401–412. 1. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. 12. Barbieri CE, Baca SC, Lawrence MS, Demichelis F, Global cancer statistics. CA: a cancer journal for clinicians. Blattner M, Theurillat JP, White TA, Stojanov P, 2011; 61:69–90. Van Allen E, Stransky N, Nickerson E, Chae SS, 2. Sakr WA, Grignon DJ, Crissman JD, Heilbrun LK, Boysen G, Auclair D, Onofrio RC, Park K, et al. Exome Cassin BJ, Pontes JJ, Haas GP. High grade prostatic sequencing identifies recurrent SPOP, FOXA1 and MED12 intraepithelial neoplasia (HGPIN) and prostatic adenocar- mutations in prostate cancer. Nature genetics. 2012; cinoma between the ages of 20-69: an autopsy study of 249 44:685–689. cases. In Vivo. 1994; 8:439–443. 13. Grasso CS, Wu YM, Robinson DR, Cao X, 3. Yin M, Bastacky S, Chandran U, Becich MJ, Dhir R. Dhanasekaran SM, Khan AP, Quist MJ, Jing X, Lonigro RJ, Prevalence of incidental prostate cancer in the general Brenner JC, Asangani IA, Ateeq B, Chun SY, Siddiqui J, population: a study of healthy organ donors. The Journal of Sam L, Anstett M, et al. The mutational landscape of urology. 2008; 179:892–895. discussion 895. lethal castration-resistant prostate cancer. Nature. 2012; 4. Weischenfeldt J, Simon R, Feuerbach L, Schlangen K, 487:239–243. Weichenhan D, Minner S, Wuttig D, Warnatz HJ, Stehr H, 14. Kluth M, Harasimowicz S, Burkhardt L, Grupp K, Rausch T, Jäger N, Gu L, Bogatyrova O, Stütz A, Claus R, Krohn A, Prien K, Gjoni J, Hass T, Galal R, Graefen M, Eils J, et al. Integrative genomic analyses reveal androgen- Haese A, Simon R, Huhne-Simon J, Koop C, Korbel J, driven somatic alteration landscape in early-onset prostate Weischenfeld J, et al. Clinical significance of different cancer. Cancer cell. 2013; 23:159–170. types of p53 gene alteration in surgically treated prostate 5. Berger MF, Lawrence MS, Demichelis F, Drier Y, cancer. International journal of cancer Journal international Cibulskis K, Sivachenko AY, Sboner A, Esgueva R, du cancer. 2014; 135:1369–1380. Pflueger D, Sougnez C, Onofrio R, Carter SL, Park K, 15. Burkhardt L, Fuchs S, Krohn A, Masser S, Mader M, Habegger L, Ambrogio L, Fennell T, et al. The genomic Kluth M, Bachmann F, Huland H, Steuber T, Graefen M, complexity of primary human prostate cancer. Nature. Schlomm T, Minner S, Sauter G, Sirma H, Simon R. 2010; 470:214–220. CHD1 Is a 5q21 Tumor Suppressor Required for ERG 6. Taylor BS, Schultz N, Hieronymus H, Gopalan A, Xiao Y, Rearrangement in Prostate Cancer. Cancer research. 2013; Carver BS, Arora VK, Kaushik P, Cerami E, Reva B, 73:2795–2805. Antipin Y, Mitsiades N, Landers T, Dolgalev I, Major JE, 16. Kluth M, Hesse J, Heinl A, Krohn A, Steurer S, Sirma H, Wilson M, et al. Integrative genomic profiling of human Simon R, Mayer PS, Schumacher U, Grupp K, Izbicki JR, prostate cancer. Cancer cell. 2010; 18:11–22. Pantel K, Dikomey E, Korbel JO, Plass C, Sauter G, et al. 7. Sun J, Liu W, Adams TS, Li X, Turner AR, Chang B, Genomic deletion of MAP3K7 at 6q12-22 is associated Kim JW, Zheng SL, Isaacs WB, Xu J. DNA copy number with early PSA recurrence in prostate cancer and absence alterations in prostate cancers: a combined analysis of of TMPRSS2:ERG fusions. Modern pathology. 2013; published CGH studies. The Prostate. 2007; 67:692–700. 26:975–983. 8. Lapointe J, Li C, Giacomini CP, Salari K, Huang S, 17. El Gammal AT, Bruchmann M, Zustin J, Isbarn H, Wang P, Ferrari M, Hernandez-Boussard T, Brooks JD, Hellwinkel OJ, Kollermann J, Sauter G, Simon R, Pollack JR. Genomic profiling reveals alternative genetic Wilczak W, Schwarz J, Bokemeyer C, Brummendorf TH, pathways of prostate tumorigenesis. Cancer research. 2007; Izbicki JR, Yekebas E, Fisch M, Huland H, et al. 67:8504–8510. Chromosome 8p deletions and 8q gains are associated with tumor progression and poor prognosis in prostate cancer. 9. Ishkanian AS, Mallof CA, Ho J, Meng A, Albert M, Clinical cancer research : an official journal of the American Syed A, van der Kwast T, Milosevic M, Yoshimoto M, Association for Cancer Research. 2010; 16:56–64. Squire JA, Lam WL, Bristow RG. High-resolution array CGH identifies novel regions of genomic alteration in 18. Koivisto PA, Schleutker J, Helin H, Ehren-van Eekelen C, intermediate-risk prostate cancer. The Prostate. 2009; Kallioniemi OP, Trapman J. Androgen receptor gene 69:1091–1100. alterations and chromosomal gains and losses in prostate carcinomas appearing during finasteride treatment for 10. Dong JT. Chromosomal deletions and tumor suppressor benign prostatic hyperplasia. Clinical cancer research : an genes in prostate cancer. Cancer metastasis reviews. 2001; official journal of the American Association for Cancer 20:173–193. Research. 1999; 5:3578–3582. 11. Krohn A, Diedler T, Burkhardt L, Mayer PS, De Silva C, 19. Alers JC, Krijtenburg PJ, Vis AN, Hoedemaeker RF, Meyer-Kornblum M, Kotschau D, Tennstedt P, Huang J, Wildhagen MF, Hop WC, van Der Kwast TT, Schroder FH, Gerhauser C, Mader M, Kurtz S, Sirma H, Saad F, Tanke HJ, van Dekken H. Molecular cytogenetic analysis Steuber T, Graefen M, et al. Genomic deletion of PTEN is of prostatic adenocarcinomas from screening studies : associated with tumor progression and early PSA recurrence www.impactjournals.com/oncotarget 27977 Oncotarget early cancers may contain aggressive genetic features. The prostate carcinoma. Journal of the National Cancer Institute. American journal of pathology. 2001; 158:399–406. 1998; 90:916–920. 20. Ribeiro FR, Henrique R, Hektoen M, Berg M, Jeronimo C, 33. Cheville JC, Lloyd RV, Sebo TJ, Cheng L, Erickson L, Teixeira MR, Lothe RA. Comparison of chromosomal Bostwick DG, Lohse CM, Wollan P. Expression of p27kip1 and array-based comparative genomic hybridization for in prostatic adenocarcinoma. Modern pathology : an official the detection of genomic imbalances in primary prostate journal of the United States and Canadian Academy of carcinomas. Molecular cancer. 2006; 5:33. Pathology, Inc. 1998; 11:324–328. 21. Jiang M, Li M, Fu X, Huang Y, Qian H, Sun R, Mao Y, 34. Erdamar S, Yang G, Harper JW, Lu X, Kattan MW, Xie Y, Li Y. Simultaneously detection of genomic and Thompson TC, Wheeler TM. Levels of expression of expression alterations in prostate cancer using cDNA p27KIP1 protein in human prostate and prostate cancer: microarray. The Prostate. 2008; 68:1496–1509. an immunohistochemical analysis. Modern pathology : an 22. Kibel AS, Faith DA, Bova GS, Isaacs WB. Loss of official journal of the United States and Canadian Academy heterozygosity at 12P12-13 in primary and metastatic prostate of Pathology, Inc. 1999; 12:751–755. adenocarcinoma. The Journal of urology. 2000; 164:192–196. 35. Vis AN, Noordzij MA, Fitoz K, Wildhagen MF, 23. Kibel AS, Schutte M, Kern SE, Isaacs WB, Bova GS. Schroder FH, van der Kwast TH. Prognostic value of Identification of 12p as a region of frequent deletion cell cycle proteins p27(kip1) and MIB-1, and the cell in advanced prostate cancer. Cancer research. 1998; adhesion protein CD44s in surgically treated patients 58:5652–5655. with prostate cancer. The Journal of urology. 2000; 24. Kawana Y, Ichikawa T, Suzuki H, Ueda T, Komiya A, 164:2156–2161. Ichikawa Y, Furuya Y, Akakura K, Igarashi T, Ito H. Loss 36. Halvorsen OJ, Haukaas SA, Akslen LA. Combined loss of heterozygosity at 7q31.1 and 12p13-12 in advanced of PTEN and p27 expression is associated with tumor prostate cancer. The Prostate. 2002; 53:60–64. cell proliferation by Ki-67 and increased risk of recurrent 25. Haugen OA, Mjolnerod O. DNA-ploidy as prognostic disease in localized prostate cancer. Clinical cancer factor in prostatic carcinoma. International journal of cancer research : an official journal of the American Association Journal international du cancer. 1990; 45:224–228. for Cancer Research. 2003; 9:1474–1479. 26. Miller J, Horsfall DJ, Marshall VR, Rao DM, Leong SY. 37. Revelos K, Petraki C, Gregorakis A, Scorilas A, The prognostic value of deoxyribonucleic acid flow Papanastasiou P, Tenta R, Koutsilieris M. p27(kip1) and cytometric analysis in stage D2 prostatic carcinoma. The Ki-67 (MIB1) immunohistochemical expression in radical Journal of urology. 1991; 145:1192–1196. prostatectomy specimens of patients with clinically localized prostate cancer. In vivo. 2005; 19:911–920. 27. Abukhdeir AM, Park BH. P21 and p27: roles in carcinogenesis and drug resistance. Expert reviews in 38. Li K, Chen MK, Situ J, Huang WT, Su ZL, He D, Gao X. molecular medicine. 2008; 10:e19. Role of co-expression of c-Myc, EZH2 and p27 in prognosis of prostate cancer patients after surgery. Chinese medical 28. Macri E, Loda M. Role of p27 in prostate carcinogenesis. journal. 2013; 126:82–87. Cancer metastasis reviews. 1998; 17:337–344. 39. Kuczyk M, Machtens S, Hradil K, Schubach J, Christian W, 29. Guo Y, Sklar GN, Borkowski A, Kyprianou N. Loss of Knuchel R, Hartmann J, Bokemeyer C, Jonas U, Serth J. the cyclin-dependent kinase inhibitor p27(Kip1) pro- Predictive value of decreased p27Kip1 protein expression for tein in human prostate cancer correlates with tumor the recurrence-free and long-term survival of prostate cancer grade. Clinical cancer research : an official journal of patients. British journal of cancer. 1999; 81:1052–1058. the American Association for Cancer Research. 1997; 3:2269–2274. 40. Wu TT, Wang JS, Jiaan BP, Yu CC, Tsai JY, Lin JT, Huang JK. Role of p21(WAF1) and p27(KIP1) in predicting 30. Tsihlias J, Kapusta LR, DeBoer G, Morava-Protzner I, biochemical recurrence for organ-confined prostate adeno- Zbieranowski I, Bhattacharya N, Catzavelos GC, carcinoma. Journal of the Chinese Medical Association : Klotz LH, Slingerland JM. Loss of cyclin-dependent kinase JCMA. 2007; 70:11–15. inhibitor p27Kip1 is a novel prognostic factor in localized human prostate adenocarcinoma. Cancer research. 1998; 41. Vlachostergios PJ, Karasavvidou F, Kakkas G, Kapatou K, 58:542–548. Gioulbasanis I, Daliani DD, Moutzouris G, Papandreou CN. Lack of prognostic significance of p16 and p27 after radical 31. Yang RM, Naitoh J, Murphy M, Wang HJ, Phillipson J, prostatectomy in hormone-naive prostate cancer. Journal of deKernion JB, Loda M, Reiter RE. Low p27 expression negative results in biomedicine. 2012; 11:2. predicts poor disease-free survival in patients with prostate cancer. The Journal of urology. 1998; 159:941–945. 42. Cordon-Cardo C, Koff A, Drobnjak M, Capodieci P, Osman I, Millard SS, Gaudin PB, Fazzari M, Zhang ZF, 32. Cote RJ, Shi Y, Groshen S, Feng AC, Cordon-Cardo C, Massague J, Scher HI. Distinct altered patterns of Skinner D, Lieskovosky G. Association of p27Kip1 levels p27KIP1 gene expression in benign prostatic hyperplasia with recurrence and survival in patients with stage C

www.impactjournals.com/oncotarget 27978 Oncotarget and prostatic carcinoma. Journal of the National Cancer Kallioniemi O-P. Survey of Gene Amplifications during Institute. 1998; 90:1284–1291. Prostate Cancer Progression by High-Throughput 43. Nassif AE, Tambara Filho R. Immunohistochemistry Fluorescence in Situ Hybridization on Tissue Microarrays. expression of tumor markers CD34 and P27 as a prognostic Cancer research. 1999; 59:803–806. factor of clinically localized prostate adenocarcinoma after 53. Minner S, Enodien M, Sirma H, Luebke AM, Krohn A, radical prostatectomy. Revista do Colegio Brasileiro de Mayer PS, Simon R, Tennstedt P, Muller J, Scholz L, Cirurgioes. 2010; 37:338–344. Brase JC, Liu AY, Schluter H, Pantel K, Schumacher 44. Sirma H, Broemel M, Stumm L, Tsourlakis T, Steurer S, U, Bokemeyer C, et al. ERG Status Is Unrelated to PSA Tennstedt P, Salomon G, Michl U, Haese A, Simon R, Recurrence in Radically Operated Prostate Cancer in Sauter G, Schlomm T, Minner S. Loss of CDKN1B/ the Absence of Antihormonal Therapy. Clinical cancer p27Kip1 expression is associated with ERG fusion-negative research. 2011; 17:5878–5888. prostate cancer, but is unrelated to patient prognosis. 54. Minner S, Jessen B, Stiedenroth L, Burandt E, Oncology letters. 2013; 6:1245–1252. Kollermann J, Mirlacher M, Erbersdobler A, Eichelberg 45. Xue W, Kitzing T, Roessler S, Zuber J, Krasnitz A, C, Fisch M, Brummendorf TH, Bokemeyer C, Simon R, Schultz N, Revill K, Weissmueller S, Rappaport AR, Steuber T, Graefen M, Huland H, Sauter G, et al. Low Simon J, Zhang J, Luo W, Hicks J, Zender L, Wang XW, level HER2 overexpression is associated with rapid Powers S, et al. A cluster of cooperating tumor-suppressor tumor cell proliferation and poor prognosis in prostate gene candidates in chromosomal deletions. Proceedings of cancer. Clinical cancer research : an official journal of the National Academy of Sciences of the United States of the American Association for Cancer Research. 2010; America. 2012; 109:8212–8217. 16:1553–1560. 46. Wang HX, Li Q, Sharma C, Knoblich K, Hemler ME. 55. Krohn A, Seidel A, Burkhardt L, Bachmann F, Mader M, Tetraspanin protein contributions to cancer. Biochemical Grupp K, Eichenauer T, Becker A, Adam M, Graefen M, Society transactions. 2011; 39:547–552. Huland H, Kurtz S, Steurer S, Tsourlakis MC, Minner S, 47. Guerillon C, Bigot N, Pedeux R. The ING tumor suppressor Michl U, et al. Recurrent deletion of 3p13 targets multiple genes: status in human tumors. Cancer letters. 2014; tumour suppressor genes and defines a distinct subgroup 345:1–16. of aggressive ERG fusion-positive prostate cancers. The Journal of pathology. 2013; 231:130–141. 48. Demichelis F, Setlur SR, Beroukhim R, Perner S, Korbel JO, Lafargue CJ, Pflueger D, Pina C, Hofer MD, 56. Mao X, Boyd LK, Yanez-Munoz RJ, Chaplin T, Sboner A, Svensson MA, Rickman DS, Urban A, Snyder M, Xue L, Lin D, Shan L, Berney DM, Young BD, Lu YJ. Meyerson M, Lee C, et al. Distinct genomic aberrations Chromosome rearrangement associated inactivation of associated with ERG rearranged prostate cancer. Genes, tumour suppressor genes in prostate cancer. American chromosomes & cancer. 2009; 48:366–380. journal of cancer research. 2011; 1:604–617. 49. Paller CJ, Antonarakis ES. Cabazitaxel: a novel second-line 57. Huang S, Gulzar ZG, Salari K, Lapointe J, Brooks JD, treatment for metastatic castration-resistant prostate cancer. Pollack JR. Recurrent deletion of CHD1 in prostate can- Drug Des Devel Ther. 2011; 5:117–124. cer with relevance to cell invasiveness. Oncogene. 2012; 31:4164–4170. 50. Erbersdobler A, Fritz H, Schnoger S, Graefen M, Hammerer P, Huland H, Henke RP. Tumour grade, 58. Mader M, Simon R, Kurtz S. FISH Oracle 2: a web server proliferation, apoptosis, microvessel density, p53, and bcl-2 for integrative visualization of genomic data in cancer in prostate cancers: differences between tumours located in research. J Clin Bioinforma. 2014; 4:5. the transition zone and in the peripheral zone. European 59. Mader M, Simon R, Steinbiss S, Kurtz S. FISH Oracle: urology. 2002; 41:40–46. a web server for flexible visualization of DNA copy 51. Mirlacher M, Simon R. Recipient block TMA technique. number data in a genomic context. Journal of clinical Methods Mol Biol. 2010; 664:37–44. bioinformatics. 2011; 1:20. 52. Bubendorf L, Kononen J, Koivisto P, Schraml P, Moch H, Gasser TC, Willi N, Mihatsch MJ, Sauter G,